The error bounds shown in the figure labeled "The Problem With Prophecy" are almost comically narrow. DTRA is very familiar with propagating uncertainty in the

The error bounds shown in the figure labeled "The Problem With Prophecy" are almost comically narrow. DTRA is very familiar with propagating uncertainty in their software. It's hard to imagine they weren't aware of how shaky these predictions were likely to be. Yet DTRA is also very familiar with the need to base decisions on the best information available at the time the decision has to be made. It's hard to imagine they weren't happy to have even shaky predictions at their disposal. These guys live in a very different world to most of us. At one point my research group was developing rules of thumb for first responders in the event of a release of a chemical/biological agent in or near a building. We vacillated on certain aspects of the advice, because no matter what guidelines we proposed, we could always think of a scenario in which that was the worst possible thing for a first responder to do. Our indecision finally got broken when a group member talked to a professional trainer of police. He flat-out stated (in paraphrase) "You don't have to be right 100% of the time. If you are right 51% of the time, that's better than having the guys on the ground making things up as they go along."